Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author.
.
u [
In the name of GOD the most beneficent the most merciful
-
-
-
n I
-
QUANTITATIVE GENETICS OF PROSTRATENESS AND OTHER RELATED ATTRIBUTES IN
RED CLOVER (Trifolium pratense L.)
A thesis presented in partial fulfilment of the requirements for the degree of
Doctor of Philosophy i n Plant Breedin g and Genetics
Department of Plant Science Massey University
New Zealand
HOSSEIN MIRZAIE-NODOUSHAN 1993
ABSTRACT
Three major experiments were conducted to investigate quantitative genetic aspects of prostrateness and related attributes in red clover (Trifolium pratense L.) during the years 1 99 1 - 1 993. These were done on several red clover genotypes with prostrate growth habit, nodal rooting ability, and early flowering characteristics, together with several other genotypes from se m ierect and erect types.
Three types of experiments were carried out:
1 ) Since genotype environment interaction is believed to be ubiquitous in affecting the performance of plants, a series of experiments were carried out in order to get general information on a range of red clover germplasm representative of the three distinct types of red clover. Twelve genotypes (four per type) were studied in a randomized complete block design with three replications at two sites for two successive years. Several techniques of univariate and multivariate analysis were applied in order to quantify and qualify the magnitude and pattern of the possible genotype-environment interaction effects. Phenotypic and genotypic correlation values were estimated for each year and type separately as well as for the whole data set in genotype-environment interaction experiment.
As a result of GE interaction analysis, a large amount of genetic variation was found in the genotypes examined. Several attributes presented significant first and second order interaction effects. 'Multivariate discriminant analysis based on these effects revealed discriminant scores by which the I contribution and importance of each attribute in the response of genotypes I examined in the environments was studied. Cluster analysis revealed that each of the three red clover types have their own particular responses to the environment effect. Phenotypic and genotypic -correlation patterns were different from year to year and type to type. Prostrate growth habit r�duced dry matter yield through significant negative correlation with yield components.
2) One accession from each of the two extreme types, erect and prostrate, were examined using a hierarchical mating design to investigate their genetic structure and to obtain more detailed genetic information on a narrower germplasm. Nine random plants from each type were cloned and used as male parent. Each male parent was crossed to six different random plants as female parents, three from the same population and three from opposite population. In other words four sets of crosses, two intra- and two inter-population sets, were IT.lade. To evaluate the 1 08 progeny families produced, male groups were divided into six sets, each containing three male groups from the same type. Each set was examined in a randomized
11
complete block design with three replications. Various genetic parameters including genetic variance components and heritability of several morphological attributes were estimated.
The two plant populations examined by the mating design, presented different patterns of genetic variation. Although the prostrate population did not have much genetic variation, its additive genetic variance components were of more importance than dominant components. However, in the erect population, dominance components of variance were more important than additive. In inter-population crosses, additive components were more important than dominance components. Stem length, number of internodes, number of branches, and plant diameter presented high level of heterosis. Number of stems, plant height, and stem thickness presented fairly high hybrid depression (negative heterosis).
Heritability broad sense and narrow sense were estimated in genotype
environment interaction experiment a·nd hierarchical mating design. Heritability values in GE interaction experiment were different from the heritability broad sense values in hierarchical mating design for most of the attributes, indicating the influence of GE interaction effect.· This difference was not noticeable in prostrateness. Heritability narrow sense estimated in hierarchical mating design varied from intra- to inter-population crosses.
3) Three sets of generation mean analysis were carried out to obtain the most detailed genetic information including function of genes, and number of genes controlling the attributes. To achieve these, three pairs of parent plants were us�d (one erect and one prostrate in each pair) to produce F1, F2, Bc1, I and Bc2• �Several attributes which were distinct enough in the two types so that it could be assumed that parent populations were nearly homozygous in opposite directions, were studied in these crosses. Three, six, and the best parsimonious models were presented for the studied attributes.
Prostrateness and stem thickness were partially to completely dominant over erectness and stem thinness. Small leaf size was over-dominant over large leaf size. There were strong evidences for additive x additive non-allelic interaction for stem thickness, additive x dominance interaction for leaf size, and dominance x dominance interaction for prostrateness and leaf size. Nodal rooting ability, prostrateness, and stem thickness seemed to be controlled by a low number of genes, whereas leaf size seemed to be controlled by several genes.
Ill ACKNOWLEDGEMENT
First of all thanks be to Almighty God the Lord of the Universe Who guides us to the right path.
Then I would like to express my particular thanks and appreciation to my sponsor, Ministry of Jahad Sazandegi of Islamic Republic of Iran for enabling me to extend my studies towards PhD by providing financial support.
I would like to express my special thanks to my chief supervisor, Dr.
I .L. Gordon. I can perhaps never thank him enough for his guidance, close assistance and very useful suggestions throughout this study.
My thanks are also extended to my eo-supervisor Dr. W. Rumball at Grassland, Agresearch, Palmerston North, for his useful and friendly suggestions and also for providing required seed.
lt was a very delightful occasion working with such a wise and excellent staff of Plant Science Department. In particular, my especial thanks are extended to Dr. K. Harrington for his guidance on weed controlling and Mrs. Karen Hill for her valuable helps in seed germinating tests.
I am very grateful and indebted to professor J. Hodgson and Professor M. Hill for their guidance, warm encouragement, and pleasant interactions.
Thanks to Professor P.G. Fenemore for pest identification and Dr. P. Long for disease verification.
I am deeply grateful to Mr. Ray Johnstone and his excellent colleagues in Plant Growth Unit (PGU). The way they handle problems is far beyond their duty.
Many thanks to field staff Mr. Terry Lynch and his colleagues for field technical assistance, F. Brown, C. McKenzie and D. Sollit for technical •.
assistance, J. Cave and C. Gwynne for providing assistance at various ways.
Finally I would iike to thank my wife, and children for their moral support, encouragement and tolerance.
IV
Table of contents
.ABSTRACT . . . . . . .-. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
_'ACKNOWLEDGEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ill
Table of contents . . . . IV
List of Tables . . . IX
List of Plates . . . ; . XIII
List of Rgures . . . . XIII
List of Appendices . . . XIV
CHAPTER ONE : INTRODUCTION
1 . 1 Introduction . . . . 1
1 .2 Overview of experimental programme . . . 4
CHAPTER TWO : LITERATURE REVIEW 2.1 Introduction . . . . 6
2.2 Genotype-environment interactions . . . . . . . . . · . . . 7
2.2.1 Analysis of variance . . . 1 0 2.2.2 Partitioning of GE interaction variance component . . . 1 3 2.2.3 Linear regression . . .· . . . . . . .·. . . . 1 4 2.2.4 Other methods of investigating GE interaction . . . 1 6 2.2.5 Multivariate methods . . . 1 7 2.2.5.1 Classification methods . . . -. . . 1 8 2.2.5.2 Principal component analysis . . . 1 9 2.2.5.3 Multiple discriminant analysis . . . . 20
2.2.6 Genotype-environment interaction in red clover . . . . 22
2.2. 7 Heritability . . . 23
2.2.8 Phenotypic and genotypic correlation . . . 26
2.3 Partitioning genetic variance . . . 27
2.3.1 Covariance of r�latives . . . 28
2.3.2 Diallel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .·. . . . . 29
V
2.3.3 Factorial mating design (North Carolina model 11) . . . 33
2.3.4 Hierarchical mating design (North Carolina model I) . . . . 35
2.3.5 Genetic variance components in red clover . . . 42
2.4 Generation mean and variance analyses . . . 43
2.5 Estimating the number of genes . . . 49
CHAPTER THREE: MATERIALS AND M ETHODS 3.1 Genotypes . . . 52
3.2 Objectives . . . 53
3.3 Genotype-environment interactions . . . 53
3.3.1 Test locations and land preparation . . . 53
3.3.2 Evaluation . . . 54
3.3.2.1 Seedling raising and Experimental design . . . 54
3.3.2.2 Studied attributes . . . . . . . . . . . . . .· . . . . 55
3.3.2.3 Statistical analysis . . . 57
3.3.2.3.1 Univariate analysis . . . 57
3.3.2.3.2 Genetical analysis . . . 61
3.3.2.3.3 Phenotypic and genotypic correlation . . 61
3.3.2.3.4 Heritability and genetic adyance. . . 62
3.3.2.4 Multivariate analysis . . . 63
3.3.2.4.1 Multivariate discriminant analysis . . . 63
3.3.2.4.2 Cluster analysis . . . 65
3.4 Hierarchical mating design . . . 67
3.4. 1 Parent plants . . . 67
3.4.2 Cloning . . . ·. 67
3.4.3 Crossing . . . · . . . 69
3.4.3.1 Field crossing nursery . . . 69
3.4.3.2 Hand crossing . . . 69
3.4.4 Evaluation . . . 70
3.4.4.1 Field design, and experimental establishment . . . 70
3.4.4.2-Studied attributes . . . 71
3.4.4.3 Data analysi� . . . 73
3.4.4.3.1 Statistical model . . . 73
VI
3.4.4.3.2 Estimating standard errors. . . . 76
3.4.4.3.3 Heritability and genetic advance . . . 77
3.4.4.3.4 Reciprocal crossing effects . . . 79
3.4.4.3.5 Heterosis and hybrid depression . . . 79
3.5 Generation mean analysis . . . . . . . . 79
3.5. 1 Crossing nursery . . . 79
3.5.2 Progeny tests . . . 80
3.5.3 Statistical analysis . . . 8 1 3.5.4 Genetical analysis . . . 8 1 3 .5.4. 1 Function of gene . . . 8 1 3 .5.4.2 Testing the model . . . 83
3.6 Estimating minimum number of genes . . . 84
3. 7 Estimating heterosis . . . 85
CHAPTER FOUR: G ENOTYPE-ENVIRONMENT INTERACTIONS RESULTS AND ASSOCIATED DISCUSSION 4. 1 Univariate analysis . . . 86
4. 1 . 1 Analysis of variance . . . . .. . . . ; . . . 86
4. 1 . 1 . 1 Environmental variances . . . 87
4. 1 . 1 .2 Genotypic variances . . . 88
4. 1 .2 Genotypic performance . . . 92
4. 1 .2. 1 Grand mean values . . . 92
4. 1 .2.2 Environments and their interaction means . . . 92
4. 1 .3 Heritability and genetic advance . . . 95
4. 1 .4 Phenotypic and genotypic correlation . . . 96 4.2 Multivariate analysis . . . 1 02 4.2.1 Multivariate discriminant analysis . . . 1 02 4.2. 1 . 1 Based on GE interaction effect (all traits) . . . 1 02 4.2. 1 .2 Based on GE interaction effect (traits with
significant GxSxY effect) . . . 1 05
4.2. 1 .3 Based _on first order interaction {Traits with
significant GxY effect) . . . 1 07
VII 4.2.2 Cluster analysis . . . 1 08 4.2.3 Type discrimination . . . ; . . . 1 1 6
CHAPTER FIVE : HIERARCHICAL MATING DESIGN
'RESULTS AND ASSOCIATED DISCUSSION
5.1 Intra- and inter-crosses mean values 1 1 8
5.2 Heterosis and hybrid depression . . . 1 18 5.3 Biometrical components of variance . . . 1 1 9
5.4 Genetic components of variance . . . 125
5.5 Dominance ratio . . . 125
5.6 Heritability and genetic advance . . . . . . . · . . . . 129
5.7 The ratio of intra- to inter-population male and female components of variance . . . � . . . 129
CHAPTER SIX : GENERATION MEAN ANALYSIS RESULTS AND ASSOCIATED DISCUSSION 6.1 Gene function . . . , . . . 136
.6. 1 . 1 Prostrateness . . . 136
kI . 1 .2 Nodal rooting ability . . . 1 39 6.1 .3 Leaf size . . . � . . . 1 41 6. 1 .4 Stem thickness . . . 142
6.2 Minimum number of genes . . . 146
CHAPTER SEVEN: GENERAL DISCUSSION --- 7.1 Introduction . . . 1 47 7.2 Genotype-environment interaction . . . 147
7.2.1 Genetic variation . . . 149
7.2.2 Environments and GE interaction . . . 149
7.2.3 Phenotypic and genotypic correlation . . . . 151
7;2.4 Multivariate analysis . . . 1 54 7.2.4.1 Multiple discriminant analysis . . . 155
VIII
7.2.4.2 Cluster analysis . . . 1 56 7.3 Hierarchical mating design . . . 1 58 7.3. 1 Why hierarchical mating design? . . . 1 58 7.3.2 Additive component of genetic variance . . . 1 5 9 7.3.3 Non-additive component of genetic variance . . . 1 61 7.3.4 Degree of dominance . . . 1 63 7.3.5 Frequency of favourable alleles . . . 1 65 7.3.6 H�ritability . . . 1 66 7.3.7 Predicted genetic advance . . . 1 68 7.3.8 Heterosis and hybrid depression . . . 1 70 7.4 Generation mean analysis . . . 1 71
7.4. 1 Introduction . . . 1 71 7 .4.2 Gene effects . . . 1 73 7.5 Number of genes . . . . . . . 1 75 7.6 Conclusions . . . 1 77 7.7 Suggested further studies . . . 1 80 REFERENCES . . . ; . . . 1 81 APPENDICES
i 1 98
IX
List of Tables
Table 2 . 1 : analysis of variance for a series of experiments pooled over
several sites and years. . . . 1 2
Table 2 .2 : Analysis of variance appropriate for design 11. . . . 34
Table 2 .3 : Coefficients of genetic components of variance . . . 35
Table 2 .4: Analysis of variance of hierarchical mating design . . . 37
Table 2 .5 : Genetic expectations of various components of variance. . . 39 Table 2 .6. Analysis of variance appropriate with design 1 when the
male groups are divided to s sets. . . . 39
Table 3. 1 : The expectations of mean squares and the degree of freedom for the analysi$ of variance of a series of experiments pooled over sites nested within year given all the effects are
random . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
'
Table 3.2: Measured attributes, their abbreviations and unit of
measurement. . . . 73
Table 3.3 : Expectations of mean square for a random effect, balance,
hierarchical mating design. . . . . 74
Table 4. 1 : The abbreviations used to refer to the attributes, the units of measurement, the grand mean, their coefficient of variance, and the values for correlation between time (r,) on which GE
intera.ction experiments were conducted. . . . 87
Table 4.2: Variance components (and their standard errors) for pooled analysis of variance of twelve genotypes at two sites over two - '
years .. . . . . . . . . . . . . . . . . . . . 90 Table 4.2, continued : Variance components (and their standard errors)
for pooled analysis of variance of twelve genotypes at two sites
over two years. . . . 9 1
Table 4.3 : Site by year overall means (environment effect) for the
measured characters. . . . 93
Table 4.4 : Genotypic means across all environments. . . . . 94
Table 4.5 : Site and year overall means and their F values for
measured characters. . . . 95
X Table 4.6: Phenotypic correlation between traits for twelve genotypes
at two sites over two years. . . . 99
Table 4.7 : Phenotypic correlations between traits for twelve genotypes at two sites for first and second year. The correlations for first year are above the diagonal and those for second year are
below the diagonal. . . . 1 00
Table 4.8 : Genotypic correlations between traits for 1 2 genotypes at two sites over two years . . .......... ..... ....... . . 1 0 1 Table 4.9 : Full and restricted heritability and their standard error. . .. . 1 02
Table 4.1 0 : Multiple discriminant of GxSxY patterns (discrimination based on all attributes) . . . ...· ... .. 1 03 Table 4.1 1 : First and second canonical structure values describing the
most part of variation exist in data set (discrimination based on all attributes). . ... 1 04
Table 4.1 2 : Multiple discriminant of GxSxY patterns in attributes with significant GxSxY effect in pooled analysis. . ... 1 05
Table 4.1 3: First and second canonical structure values describing the most part of variation exist in data set (discrimination based on attributes with significant GxSxY effect in ANOVA). , ... 1 06
Table 4J 4 : Multiple discriminant of GxY patterns in attributes with
si
�
nificant GxY effect in pooled analysis. . ... 1 07Table 4.1 5: The structures of discriminant functions and standardized coefficients of the analysis based on the attributes with significant GxY effect. ... 1 08
Table 4.1 6 : The probability of F test for different clustering stages · (clustering based on GxSxY effect and all attributes). . ... 1 0 9
Table 4.1 7 : The probability of F test for different clustering stages
(clustering based _on attributes with significant GxSxY effect) .. � . 1 1 0 Table 4.1 8 : The probability of F test for different clustering stages
(clustering based on attributes with significant GxY effect). . ... 1 1 0
Table 4.1 9 : Cluster mean values for all environments (clustering based
on GxSxY effect) . . . . 1 1 1
_ Table 4.20: Cluster mean values for both years (clustering based on
XI attributes with significant GxY effect). . ... 1 1 2
Table 4.2 1 : Multiple discriminant analysis results based on the growth
habit. ... � . . . . 1 1 6
Table 4.22 : First ·and second canonical structure values and
standardized canonical structures. . ... 1 1 7
Table 5.1 : Mean values, their percentage of coefficient of variation (in brackets), heterosis, hybrid depression for inter and intra
population sets of crosses (F1). • • • • • • • • • • • • • • • • • • • • • • • • 1 20 Table 5.2 : Biometrical components of variance and their standard
errors in erect intra-population crosses. . ... 1 2 1
Table 5.3 : Biometrical components of variance and their standard · errors in prostrate intra-population crosses ... 1 22
Table 5.4 : Biometrical components of variance and their standard errors in first inter-population crosses where erect population
was as male parent population. . ... 1 23
Table 5.5 : Biometrical components of variance and their standard errors in second inter-population crosses where prostrate
population was as male parent population. . ... 1 24
Table 5.6 : Estimates of genetic variance components and their
1 standard errors in erect population. . ... 1 26
Ta\ �le 5.7 : Estimates of genetic variance components and -their standard errors in prostrate population. : ... 1 27
Table 5.8 :Estimates of genetic variance co�ponents and their standard errors in inter-population crosses ... 1 28
Table 5.9 : Full heritability narrow sense and broad sense estimates - .
and expected geneti<? advance per cycle of selection in intra and
inter-population crosses. . ... 1 30
Table 5.1 0 : Restricted heritability narrow sense and broad sense estimates and expected genetic advance per cycle of selection '
in intra and inter-population crosses. . ... 1 31
Table 5.1 1 : Ratios of intra to inter-population male and female components of variances. . ... 1 32
Table 6.1 : Degrees of freedom and mean squares from the weighted
XII analyses of variance of parental, F11 F2, 81, and 82 for 4
characters in three sets of crosses (erect x prostrate) . . . . . . . . . 1 34 Table 6.2 : Observed generation means, their within plot variances and
F1 mid-parent deviations (F1-MP) for four attributes in cross
one . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 34 Table 6.3 : Observed generation means, their within plot variances and
F1 mid-parent deviations {F1-MP) for four attributes in cross
two . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 35 Table 6.4 : Observed generation means, their within plot variances and
F1 mid-parent deviations (F1-MP) for four attributes in cross
three . . . 1 35
Table 6.5: Gene effects estimated for prostrateness using three and six parameter models on mea'ns and their variances of parents, F1, F2, 81 and 82 in a cross between erect and prostrate
plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 38 Table 6.6 : Gene effects estimated for nodal rooting ability using
three and six parameter models on means and their variances of parents, F1, F2, 8c1 and 8c2 in a cross between erect and
prostrate plants. . . . . . , . . . 1 40
Table 6 .. 7 : Gene effects estimated for le�lf size using three and six p�I rameter models on means and their variances of parents, F1,
F2; Bc1 and 8c2 in a cross between erect and prostrate plants ' . . . 1 44
Table 6.8 : Gene effects estimated for stem thickness using three and - six parameter models on means and their variances of parents, F1, F2, 8c1 and 8c2 in a cross between erect and prostrate - plants . . . 1 45
Table 6.9 : The estimated number of genes control the attributes (and
their standard errors). . . . . 1 46
XIII
List of Plates
Plate 1 : General performance of the two types of red clover, prostrate
and erect . . .... ; . . . 66
Plate 2: General view of plots after transplanting the jiffy pots in the
field ................ . ..... . . ... . . . . .... . ..... 70 Plate 3 : Variation in leafiness in inter-population crosses. . . . . 72
List of Figures
Figure 4. 1 : Dendrogram for cluster analysis based on the first and second discriminant scores for the GxSxY partition (all
attributes). . . . 1 1 3
Figure 4.2 : Dendrogram for cluster analysis based on the first and second discriminant scores for the GxSxY partitioning (attributes
with significant GxSxY effect in ANOV A). . . . 1 1 4
Figure 4.3 : Dendrogram for cluster analysis based on the first and second discriminant scores for the GxY partitioning (attributes
with significant GxY effect in ANOVA). . . . 1 1 5
XIV
List of Appendices
Appendix 1 , Table 1 : accession number and their structure used in
experiments. . . . . 1 98
Appendix 1 , Table 2: The results of the test of homogeneity of
variance in the genotype-environment interaction experiments . . . 1 99
Appendix 2, Table 1 : Genotype X site interaction mean values. . . . 200 Appendix 2, Table 1 , continued : Genotype X site interaction mean
values. . . . 201
Appendix 2, Table 2 : Genotype X year interaction mean values . . . 202 Appendix 2, Table 2, continued : Genotype X year interaction mean
values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Appendix 2, Table 3 : Genotype X environment Jnteraction mean
\ values. . . . . 204 Appendix 2, Table 3, continued: Genotype-environment interaction
mean values. . . . 205
Appendix 3 : Experimental mean squares, (their corresponding F -
values) from pooled analysis of variance of twelve genotypes at
two sites over two years. . . . . 206
Appendix 3; continued : Experimental mean squares, (their •.
corresponding F values) from pooled analysis of variance of
twelve ge-notypes at two sites over two years. . . . . 207 Appendix 4, Table 1: Phenotypic and genotypic correlation between
traits for three genotypes in erect type. The phenotypic correlations are above the diagonal and genotypic correlations
XV
are below the diagonal. . . . 208 Appendix 4, Table 2 : Phenotypic and genotypic correlation between
traits for three genotypes in semi-erect type. The phenotypic correlations are above the diagonal and genotypic correlations
are below the diagonal. . . . 209 Appendix 4, Table 3 : Phenotypic and genotypic correlation between
traits for three genotypes in prostrate type. The phenotypic correlations are above the diagonal and genotypic correlations
are below the diagonal. . . . . . . . . . . . . . . . . . . . . . . . . . . 210 Appendix 5: HOMINO programme for analysing generation means,
estimating number of genes and heritability . . . . . . . . . . . . . . 21 1
